The production of nitriles by ammoxidation of an appropriate hydrocarbon in the presence of a suitable catalyst is well known. The production of acrylonitrile, for example, from a gaseous feed of propylene, ammonia and air is described by Bruce E. Gates et al in Chemistry of Catalytic Processes, McGraw-Hill (1979), pp. 380-384.
The feed is sent to an ammoxidation reactor where, in the presence of a suitable catalyst, acrylonitrile is produced along with lesser amounts of other nitrogen containing compounds. The effluent from the ammoxidation reaction is quenched with water and the desired products are obtained in the liquid phase. The gas phase byproducts, typically oxygen, carbon dioxide, carbon monoxide and unreacted hydrocarbon, are combined with natural gas and sent to a boiler for combustion as disclosed, for example, in Yoshino et al., U.S. Pat. No. 3,591,620 and Callahan et al., U.S. Pat. No. 4,335,056.
More recently, Khoobiar et al., in U.S. Pat. No. 4,609,502 disclosed a cyclic process for producing acrylonitrile using propane as a starting material which is initially dehydrogenated catalytically in the presence of steam to form propylene. This is in contrast to most conventional dehydrogenation processes which avoid steam primarily due to the costs involved. After ammoxidation, the effluent is quenched, the desired product removed, and the off-gases, including propylene and propane, are sent to an oxidation reactor to remove oxygen by selective reaction with hydrogen to form water vapor. The gas mixture exiting the selective oxidation reactor includes substantial amounts of methane, ethane and ethylene, which are by-products of dehydrogenation, and unreacted propylene in addition to carbon oxides. As an option, this gas mixture is split and a portion is sent to a separator which removes only carbon dioxide. A portion of the effluent from the separation is purged to remove light hydrocarbons. The nonpurged stream is mixed with the remainder of the oxidator effluent, fresh propane and steam, if necessary, and sent to the dehydrogenator where the propane is converted to propylene. Another option is to cool and liquify the C.sub.3 hydrocarbons therefrom and then vaporize them prior to recycle.
The aforementioned process suffers from several disadvantages. The amount of propane which is converted to propylene is only in the range of about 20 percent to 60 percent per pass, typically about 40 percent, and therefore, about 60 percent of the propane feed is recycled throughout the system. At conventional velocities, the presence of such large amounts of propane along with hydrogen and other gases can produce higher pressures in the ammoxidation reaction zone which can, in turn, result in decreased yields of acrylonitrile. This problem could be overcome by using a more efficient dehydrogenation catalyst, if such were commercially available. Also, there is no practical way in this scheme to selectively remove by-products of propane dehydrogenation, such as methane, ethane, ethylene and the like, thereby preventing their accumulation in the system. Providing a purge stream to remove these gases will also cause removal of some of the circulating propane and propylene. As the process is being carried on in a continuous manner, this loss of starting material causes a significant decrease in the yield of propylene. It is disclosed that propane and propylene are recovered from the purge stream prior to venting. Additional refrigeration is therefore necessary to liquify the propane and propylene. This apparatus, as well as that required to vaporize them prior to recycle, significantly adds to the capital cost of the process.
Another disadvantage of the Khoobiar et al process stems from the use of the selective oxidation reactor to treat the gaseous effluent from the quencher. The gases exiting the quencher are at ambient temperature and must be heated prior to introduction into the oxidation reactor in order to promote oxygen removal. Because there is a significant amount of oxygen in the quencher effluent, the heat of reaction generated in the oxidation reactor can result in excessive temperatures in the system. There are three options to alleviate this problem. First, the amount of oxygen entering the oxidation reactor can be reduced by other means. Second, multiple reactors can be utilized with a cooling means between each pair of reactors. Third, a portion of the effluent from the reactor is passed through a cooling means and recycled to the feed to reduce the internal temperature of the reactor. None of these measures is attractive from the viewpoint of cost and efficienty.
The oxidation reactor in the Khoobiar et al process is operated with oxidation catalysts such as noble metals (e.g., platinum). Olefins and carbon monoxide, which are generated in the dehydrogenation reactor, are known to poison these catalysts, as disclosed in Catalytic Processes and Proven Catalysts, Charles L. Thomas, Academic Press (1970) pp. 118-119. Accordingly, multiple oxidation reactors must be used to allow for frequent regeneration of the catalyst which represents yet another addition to production costs (U.S. Pat. No. 4,609,502, column 4, lines 51-56).
It is therefore apparent that industry is still searching for a cost effective process of converting hydrocarbons into nitriles or anhydrides. Applicants have discovered a process which is cost effective and in which the disadvantages of the aforementioned systems are substantially reduced or eliminated. Moreover, in comparison to conventional processes, the thermal requirements of Applicants' process are markedly reduced.